idea-creator

To generate publishable research ideas from a broad ML direction, provide the research area to trigger a landscape survey that identifies gaps and outputs 8–12 ranked ideas with feasibility and novelty checks. You receive a structured Idea Report suitable for conference submissions.

A landscape survey for identifying research gaps in AI systematically maps a broad research area to pinpoint unexplored problems. It generates actionable research directions and evaluates their feasibility before moving to pilot design.

To run a novelty check and pilot design for machine learning research, configure parameters like PILOT_MAX_HOURS and MAX_TOTAL_GPU_HOURS to validate 1–2 selected ideas. The skill outputs phase-wise validation plans with evaluation metrics and pilot-ready experimental setups.

Yes, you need Codex MCP integration to use this automated academic pipeline tool, as it requires this setup alongside standard research datasets to function. Configurable parameters like MAX_PILOT_IDEAS also depend on this environment to generate pilot-ready plans.

Yes, you can configure GPU hour limits for pilot study design by setting the MAX_TOTAL_GPU_HOURS and PILOT_MAX_HOURS parameters. This constrains the computational budget for validating 1–2 research ideas, ensuring pilot experiments remain resource-efficient.

Automated research idea generation is limited by its dependency on Codex MCP integration and standard research datasets. It focuses strictly on ML/AI domains and requires manual oversight to execute pilot studies beyond the configured MAX_PILOT_IDEAS limit.